A nonprofit wishes to build a 95% confidence interval on the proportion of uninsured Americans in the United States. In 2010, 16% of Americans had no insurance. In 2016, it was 8.6%.
a) If the nonprofit prefers to estimate the sample size conservatively, which of the two historical proportions of uninsured Americans should they use?
b) Using the most conservative historical proportion and a margin of error of 1%, determine the required sample size.